This research will focus on the impact of psychosocial factors on health outcomes in the presence of a shared chronic health condition requiring standardized therapy, and how this impact may vary by gender, race, and age. The study population will consist of older adults receiving dialysis treatment because of chronic kidney failure, and a matched control group of older adults who do not share this health condition. The first objective of the study is to prospectively investigate contributors to survival of 350 older black and older white persons who were receiving maintenance kidney dialysis when they were interviewed in 1988. The analysis will investigate the contributions of psychosocial variables to survival, over and above the contributions of the variables gender, race, age, primary cause of kidney failure, number of comorbid conditions, and hospitalization history that have been identified as predictive in other studies of dialysis patient survival. The analysis will also investigate whether the relation of psychosocial variables to survival of older dialysis patients varies significantly by gender/race category. The relation of psychosocial variables to survival among older black dialysis patients has not been investigated. The Cox proportional hazards model with time dependent covariates (the psychosocial variables) will be used; assumptions will be tested. The second objective of this study is to reinterview those persons from the sample of 350 older dialysis patients interviewed in 1988 who are still alive and living in Georgia in 1991 and to reinterview persons from a sample of 340 matched community respondents (control group) who were interviewed in 1988 and are still alive and living in Georgia in 1991. Subjects' reinterview date will be three years after the date of the initial interview. The purpose of the followup is to investigate (1) stability/change in subjects' psychological and social functioning over time, (2) whether these patterns differ significantly between the dialysis and control samples, and (3) how gender, race and age may be related to observed patterns of psychological and social functioning. The general linear model or categorical data analysis (Grizzle-Starmer-Koch) will be used, with main effect terms of gender, race and age and their interaction terms testing the hypotheses of interest.